Explore

The Galapagos Hub

Scaffolds, tasks and discoveries from the community — searchable, composable, and labelled with their current review state.

4136 cards

16 scaffolds

SkyDiscover/adaevolve
Test Time Search
Hierarchical adaptive search: G-signal exploration intensity, UCB island allocation, and LLM meta-guidance on stagnation.Updated today91
SakanaAI/ale_agent
Test Time Search
Domain-guided tabu best-first search with 30 sibling branches and three-turn solution refinement.Updated today30
AlgoTune/algotune_agent
Test Time Search
Command-driven algorithm optimizer with evaluation tools, profiling, multi-file edits, and best-snapshot restore.Updated today40
default/beam_search
Test Time Search
Maintain a fixed-width beam of promising programs; expand one per step, prune by fitness+diversity.Updated today80
default/best_of_n
Test Time Search
Give the LLM N valid attempts at the same parent before committing to the global best, then repeat.Updated today90
default/best_of_n_attempts
Test Time Search
Best-of-N that rotates the parent every N attempts — failed/invalid tries spend the budget too.Updated today100
default/claude_code
Test Time Search
Single-agent baseline that hands the whole search loop to one Claude Code CLI session, run inside the task's own container image by default (subscription-only billing): Claude edits the solution and runs the task's own evaluator, and the framework scores checkpoints with that same evaluator. A run scored on the host (general.eval_mode: local) gets the session as a host subprocess instead.Updated today50
default/codex
Test Time Search
Single-agent baseline that gives one Codex CLI session the complete search loop inside the task's own image: Codex edits and evaluates, while Galapagos independently scores every observed checkpoint.Updated today20
SkyDiscover/evox
Test Time Search
Co-evolves the search strategy with the solutions: the parent/context selection policy is itself LLM-written code, scored by windowed improvement and hot-swapped on stagnation.Updated today100
GEPA/gepa
Test Time Search
Reflective mutation over execution feedback, with Pareto-frontier candidate selection.Updated today00
SkyDiscover/gepa_native
Test Time Search
Stored-parent reflective evolution with an elite pool, rejection history, and LLM-mediated merge.Updated today00
Meta-Harness/meta_harness
Test Time Search
A minimal outer loop that delegates selection AND mutation to a skill-steered proposer over an append-only candidate history, returning a scalar or task-configured Pareto frontier.Updated today90
OpenEvolve/openevolve
Test Time Search
Island-model MAP-Elites evolutionary search with diff mutation (the open AlphaEvolve).Updated today70
default/openevolve_world_knowledge
Test Time Search
OpenEvolve with an open-world retrieval layer: a gate decides when to search the web, and the search query is itself evolved.Updated today00
community-example/random
Test Time Search
Uniformly sample a parent and context from the full candidate history.Updated today00
default/topk
Test Time Search
Always expand the single best program, with the next K as context. Pure greedy elitism.Updated today80

The Galapagos Hub

SkyDiscover/adaevolve

SakanaAI/ale_agent

AlgoTune/algotune_agent

default/beam_search

default/best_of_n

default/best_of_n_attempts

default/claude_code

default/codex

SkyDiscover/evox

GEPA/gepa

SkyDiscover/gepa_native

Meta-Harness/meta_harness

OpenEvolve/openevolve

default/openevolve_world_knowledge

community-example/random

default/topk